Extend HostToDomainGraph to fold host-level graphs stripping the www. prefix

[lfoppiano]

As specified in #29

[lfoppiano]

@sebastian-nagel thanks for the comments, just to be sure, in both the pointed parts, the variable host should hold the URL in the standard way (www.google.com), because it was reverted from the input which is in the SURT? approach (com.google.com), or at least this is what I have understood from the tests

[sebastian-nagel]

host should hold the URL in the standard way (www.google.com)

Sorry, yes, of course. Also EffectiveTldFinder needs the unreversed form. That makes the code even simpler.

[lfoppiano]

Yes. AFAIK the variable host carries the unreversed form. So, anything that come after would work left to right if I haven't overlooked anything.

[sebastian-nagel]

Thanks, @lfoppiano!

Sorry, there was an conceptual misunderstanding: host-without-www is on the same level as are registered and private domain.

The three nodes

0       com.example
1       com.example.www
2       com.example.xyz

are folded to (without counts of "subdomains"):

0       com.example       2
1       com.example.xyz       1

I'll update the script src/script/host2domaingraph.sh to work with the option --aggregation-level requiring an option value.

[sebastian-nagel]

Tabs in the whitespace. If the terminal shows tabs with length 8, then the help might look skewed:

 --private-domains      (deprecated - use --aggregation-level)
                                        convert to private domains (include suffixes from the
                        PRIVATE domains subdivision of the public suffix list,
                        see https://github.com/publicsuffix/list/wiki/Format#divisions)

[sebastian-nagel]

For testing, could run as:

./src/script/host2domaingraph.sh -h 2>&1 | expand -8

[lfoppiano]

If I have understood correctly I've made all the required changes.

[sebastian-nagel]

Thanks! Looks good to me. Tested with a small host-level web graph and all three aggregation levels.

Please squash the commits to a smaller meaningful number.